Vcsn 2

Vcsn 2

Vcsn is a finite state machine manipulation platform, consisting of a templated C++ library, a dynamic C++ API, and Python GUI.

See http://vcsn.lrde.epita.fr.

Name Last Update
bin Loading commit data...
build-aux Loading commit data...
buildfarm Loading commit data...
debian Loading commit data...
debuild Loading commit data...
doc Loading commit data...
lib Loading commit data...
libexec Loading commit data...
ports Loading commit data...
python Loading commit data...
share Loading commit data...
tests Loading commit data...
vcsn Loading commit data...
.clang-format Loading commit data...
.clang-tidy Loading commit data...
.dir-locals.el Loading commit data...
.gitignore Loading commit data...
COPYING.txt Loading commit data...
Dockerfile Loading commit data...
Makefile.am Loading commit data...
NEWS.md Loading commit data...
README.md Loading commit data...
TODO.txt Loading commit data...
bootstrap Loading commit data...
configure.ac Loading commit data...
internship.txt Loading commit data...

Vcsn

Vcsn is a platform for weighted automata and rational expressions.

It consists of an efficient C++ generic library, shell tools, Python bindings, and a graphical interactive environment on top of IPython.

Copyright (C) 2012-2017 The Vaucanson Group.

Overview

The Vcsn platform enables the development of C++ programs manipulating weighted finite automata in an abstract and general way with, at the same time, a large specialization power. On the one hand, we can write algorithms working on every automaton with weights in any semiring and with words from any free monoids. And on the other hand, a particular algorithm can be specialized for a particular data structure.

The Python bindings, and especially the IPython interface, make Vcsn a tool particularly well suited for practical sessions in courses of Formal Language Theory. More generally, it proves to be a handy means to explore compositions of algorithms on automata from small sizes to "real world" cases.

Although it is now quite mature, Vcsn is an ongoing development project. Therefore some algorithms and data structures may change in the future.

Please send any question or comments to vcsn@lrde.epita.fr, and report bugs to either our issue tracker https://gitlab.lrde.epita.fr/vcsn/vcsn/issues, or via emails to vcsn@lrde.epita.fr.

Using Vcsn

Documentation about Vcsn can be found in several places:

  • the directory doc/notebooks documents the Python interface and provides examples. It is also helpful to understand the C++ API. This documentation is also available on line. The online notebooks will be frequently updated, so please be sure to check there. In particular, be sure to read the introduction to Vcsn.

  • the file NEWS.md includes many examples of how to run commands and algorithms.

  • the directory tests/python contains tons of test cases written in Python.

  • the C++ low-level interface is documented via comments in the header files (vcsn/algos/*.hh).

  • the dyn:: C++ interface is documented in vcsn/dyn/algos.hh.

Installation

To install Vcsn on your system, type in the classical sequence at the command prompt:

./configure
make
make install (as root)

Do not hesitate to run make -j3 if, for instance, your CPU features 4 threads. To enable the generation of the Doxygen documentation, pass --enable-doxygen to configure.

Note that an installation is specific to the compiler used to install it. Indeed, the call to ./configure enables some workarounds and, consequently, users must compile with the same compiler to avoid compatibility problems.

Between make and make install, you may also want to run:

make check

It run the test suite to check the whole platform. Beware that checking Vcsn is a very long process, also consider -j3.

Build Requirements

Packages needed

Vcsn was tested with the GNU Compiler Collection (GCC) versions 5, 6 and Clang 3.5, 3.6, 3.7, 3.8, 3.9.

Boost is a C++ library which provides many useful features. You must install this library on your system. Vcsn should support any version after 1.49. The following Boost components are used:

  • Boost.Algorithm
  • Boost.DynamicBitset
  • Boost.Filesystem
  • Boost.Flyweight
  • Boost.Heap
  • Boost.Iterator
  • Boost.Python
  • Boost.Range
  • Boost.Regex
  • Boost.System
  • Boost.Tokenizer

Ccache saves the user from repeated compilations.

To load plugins, Vcsn relies on libltdl, which is a component of the GNU Libtool project. Depending on your distribution/packaging system, you may have to install libltdl-dev (e.g., Debian) or libtool (Mac Ports).

Vcsn uses the Dot format to save automaton in a human readable file. You should install Graphviz to visualize these .gv files.

To provide safe support for ℚ, Vcsn relies on The GNU Multiple Precision Arithmetic Library.

Doxygen is used to generate the C++ reference documentation.

yaml-cpp is used to handle the configuration files. Beware that version 0.5.2 is buggy and will not work properly. Use 0.5.1, or 0.5.3 or more recent.

Ubuntu/Debian packages

Please, help us keep this list up-to-date!

ccache dot2tex g++ graphviz imagemagick ipython3-notebook libboost-all-dev libgmp-dev libzmq3-dev locales pdf2svg python3-colorama python3-dev python3-matplotlib python3-pandas python3-pip python3-psutil python3-regex python3-setuptools libyaml-cpp-dev

Vcsn expects to be built and to run in an UTF-8 environment. This requires the locales package, and that en_US.UTF-8 be supported:

sudo echo "en_US.UTF-8 UTF-8" >/etc/locale.gen
sudo locale-gen
export LANG=en_US.UTF-8   \
       LANGUAGE=en_US:en  \
       LC_ALL=en_US.UTF-8

Libraries installed in non-standard directories

If you have installed Boost in a non-standard directory (i.e., a directory that is not searched by default by your C++ compiler), you will have to set the CPPFLAGS and LDFLAGS variables to pass the necessary -I and -L options to the preprocessor and linker.

For instance if you installed Boost in /opt/boost/ you should run ./configure as follows:

./configure CPPFLAGS="-I/opt/boost" LDFLAGS="-L/opt/boost"

Layout of the tarball

The project directory layout is as follows:

build-aux : Auxiliary tools used by the GNU Build System during configure and make stages.

doc : Doxygen documentation, and IPython notebooks.

share : Data files to be installed on your system.

lib : Various libraries, including instantiation of some contexts.

vcsn : The Vcsn C++ Library headers.

python : The Python binding.

bin : Various programs to install. In particular the program vcsn, which provides access to all the other programs. See vcsn --help.

tests : The test suites.

Starting from the repository

To contribute to Vcsn, or to build it from its Git repository, you need more tools:

  • Automake 1.14 or newer
  • Autoconf 2.69 or newer
  • Bison 3.0.4 or newer
  • Flex 2.5.35 or newer

Before the configuration steps, run:

./bootstrap

to set up the GNU Build system.

Ubuntu

In addition of the packages above, you need:

autoconf automake libtool flex bison

License

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.

The complete GNU General Public License Notice can be found as the COPYING.txt file in the root directory.

Contacts

The team can be reached by mail at vcsn@lrde.epita.fr. Snail mail addresses follow.

  • Vcsn - LRDE

Akim Demaille & Alexandre Duret-Lutz\ Laboratoire de Recherche et Développement de l'EPITA\ 14-16 rue Voltaire\ 94276 Le Kremlin-Bicêtre CEDEX\ France